Research Highlights

Explorations: Some New Problems

Optimizing Policy in Epidemic-Human Behavior Co-evolution: Graph Algorithms to identify best policies under simple models of epidemics and behavior flow.

Better approximation algorithms for Immunization and Influence Maximization.

Data-driven Discovery of Natural Laws: Designing special sparse deep learning networks to identify equations that represent the data

Simultaneous Learning of Architecture and Parameters: Instead of fixing an architecture and training parameters can gradient descent lead to both simultaneously? [FITML@NeurIPS]

Simplicity-Complexity Gap in Epidemic Models: Given a certain amount of noise how much do simple models differ from complex models in homogenous mixing models and network models

Graph Neural Networks (Applications)

We develop graph neural networks for several applications. These include: (i) Autism detection [ICASSP 2023], (ii) Prediction of emergence of virus variants [arXiv], and (iii) Understanding networks and suicide risk [arXiv]

Infectious Disease Modeling, Forecasting, and Projections

We developed the SIkJalpha model that can capture crucial factors and variables that assist with projecting desired future scenarios. We have used this model in multi-model collaborative efforts to predict short-term outcomes (cases, deaths, and hospitalizations) of COVID-19, Influenza, and RSV and long-term scenario projections. We have participated in many such efforts: US Scenario Modeling Hub, US Forecast Hub, Europe Scenario Modeling Hub, Europe Forecast Hub, Germany/Poland Forecast Hub. For Influenza forecasting we proposed a Tree Ensemble model design that utilizes the individual predictors of SIkJalpha to improve its performance.

[Paper on evolution of the model][Scenario Modeling at CDC MMWR][PNAS on US Forecast Hub][Influenza paper][Influenza presentation]

Shape-based Representation , Evaluation, Clustering, Classification, and Ensembles

We developed a transformation of the numerical forecasts into a shapelet-space representation. We prove that this representation satisfies the property that two shapes that one would consider similar are mapped close to each other, and vice versa. We develop a shapelet-space ensemble of multiple models which is the mean of the shapelet-space representation of all the models. We show that this ensemble can accurately predict the shape of the future trend.

For long-term projections, we develop a shape-based alignment and ensembling technique that leads to preservation of properties of underlying trajectories that are often missed by traditional ensembles (such as mean peak size). We have also shown that this approach leads to better clustering and classification.

[BigData 2022 Paper:SSR][AAAI 2025 Paper:DTW+S][Presentation]

Graph Neural Networks (Theory)

Training and inference on deep GNNs on large graphs are difficult due to computational complexity and lack of accuracy improvements with deeper layers. Subgraph-based methods to address training on large graphs exist, but they do not apply during inference, making inference the bottleneck. Such methods also do not address poor accuracy for deep networks due to "oversmoothing". We address the following challenges: (i) Developing subgraph-based schemes that apply to training and inference. (ii) Identifying good subgraph-sampling strategies. (iii) Pruning weights to reduce computations during inference.

NeurIPS 2021, VLDB 2021, ICLR 2020, IPDPS 2019

Past Research

Prior to my faculty position, I worked on a range of problems spanning from theoretical to experimental to real-world deployments, that involved a mix of Algorithms, Network Science, and Data Mining. The figure summarizes my past research. Please see my Publications page or contact me to learn more about my contributions to these problems.

Page updated

Google Sites

Report abuse